Table of contents

Recent attempts at a microscopic description of the liquid-solid transition, within the density-functional theory of non-uniform fluids, are put into a critical perspective by comparing the different methods proposed and the results obtained for the freezing of hardcore systems.

The magnetic behaviour of highly magnetostrictive amorphous wires is a direct consequence of the residual stresses developed during the sample fabrication. A phenomenological model that assumes a heat transference from a metallic liquid to the water, with cylindrical symmetry, and a solidification process that propagates in the opposite direction to the heat flow has been proposed. The value for the tensile radial stress field has then been calculated. A maximum value for these radial stresses has been found at a distance of about 0.7 R from the axis of the wire.

The magnetic susceptibility was measured in plastically deformed 25.6 at.% Fe-Pt single crystals. The susceptibility decreases considerably as a result of a small strain epsilon of 1.3% and at the same time spontaneous magnetisation appears. The spontaneous magnetisation increases with increasing plastic deformation. On cooling, the spontaneous magnetisation increases rapidly near the Neel temperatures and has a local maximum at 150 K. The dislocation density and its distribution were also observed by transmission electron microscopy. Superlattice dislocations are distributed making a pair. Near the anti-phase boundary between two superpartials, Fe atoms occupy the face-centred sites and couple ferromagnetically. The relation between the spontaneous magnetisation and the dislocation density is discussed using a simple localised moment model. It is shown that the ferromagnetic transition is extended as far as about twentieth nearest neighbour from the anti-phase boundary. The influence of the ferromagnetic clusters on their neighbouring Fe atoms is extended as far as 10² nm and the Fe atoms change to being superparamagnetic. This long-distance influence and the local maximum of spontaneous magnetisation should be explained from the viewpoint of the itinerant-electron model.

In modelling ionic crystals with the rigid-ion model under the adiabatic and harmonic approximations, the series involving the long-range Coulomb forces appear to be only conditionally convergent. The approach most commonly used to sum these series is the Ewald method, which separates them into two rapidly convergent parts in real and reciprocal space. Instead, the technique described here makes use of infinite sheets of charge to cancel effectively the long-range moments of these forces. The original series can then be rewritten in a form which rapidly converges, but still only in real space. The new method was tested by successfully duplicating Kellermann's model (1940) results, which made use of the Ewald method. The new technique is restricted to lattice dynamical problems as opposed to static problems but may be applied to any crystal lattice and incorporated into any lattice dynamical model which also makes use of the harmonic approximation.

Earlier experimental results have been considerably strengthened by Raman scattering from crystalline sym-C₆Cl₃F₃. Measurements have been made as a function of temperature at ambient pressure, and as a function of pressure at room temperature. Apart from a new first-order phase transition observed at 62 kbar but as yet uncharacterised, the crystal structure maintains the P6₃/m space group symmetry. Anomalies in phonon frequencies and line-widths at 296 K indicate a phase transition, which is 'isostructural' order-disorder. Similar features are seen at 15 kbar, suggesting that the phase transition is of the same nature as that observed at 296 K. The half-width of the phonon associated with dynamic disorder is found by fitting to be Gamma ₁2/(T)=0.032+0.0061 T+3400 exp(- Delta E_a/kT)cm^-1, with Delta E_a=35 kJ mol^-1. Although in-plane 120 degrees molecular reorientational jumps may play a role, this may not be a major role at room temperature due to a rather high activation energy. The full understanding of the results requires there to be another minor activation process that also obeys A_g symmetry. The variation in the crystal structure parameters and the external mode frequencies with pressure up to 60 kbar is calculated using a 6-exp atom-atom potential with the rigid-body approximation. The molecular orientation changes by only 0.5 degrees between 0 and 60 kbar. The calculated external mode frequencies agree with Raman experiment up to 60 kbar within 5-15%. A molecular dynamics simulation for 300 K indicates a residence time between reorientations greater than 500 ps, giving a quasi-elastic Lorentzian broadening of less than 0.01 cm^-1.

According to small-angle X-ray scattering, amorphous Ni-P contains some volume per cent of metal-rich precipitations. During annealing, their diameters grow from about 2 nm in the as-prepared state to about 6 nm. It is shown by magnetic measurements that the inclusions are ferromagnetic single-domain particles embedded in a paramagnetic matrix. At sufficiently high temperatures such systems behave like superparamagnets. At 4.2 K, thermal agitation is too weak to flip the elementary magnetic moments. In this case, magnetisation curves can allow the authors to determine the coercivities and average saturation magnetisations. Annealing induces the two quantities measured at 4.2 K to change proportional to each other. This experimental fact justifies the application of a theory developed for fine non-interacting ferromagnetic particles having a magnetic shape anisotropy. The increase in the spontaneous magnetisation is attributed to the decomposition of the inclusions. With the aid of an approximation procedure the starting value of the phosphorus content and its decrease during annealing are estimated. These results and the values derived for the shape anisotropy agree well with those obtained through small-angle X-ray scattering. The starting phosphorus concentration of the precipitations proved to be independent of their volume fraction and the average composition of the samples. The observed kinetics fit in with the assumption of an activation energy spectrum. The lower limit starts at about 1.4 eV and is shifted on annealing to higher values.

Similar to the case of the spin-lattice coupling coefficients G_ij in cubic symmetry, in this paper the simple correlations between the coupling coefficients F_ij(F₁₁, F₁₂ and F₄₄) in a cubic field and the g-factor in cubic and low symmetries have been established, and hence a simple and uniform method suitable for calculating the coefficients F_ij for all dⁿ ions is given. As an example, the analytic expressions for F_ij for d³ ions in a cubic field are obtained from the high-order perturbation formulae for the g-factor. From these expressions, the coefficients F_ij for MgO:Cr³⁺ crystals have been calculated. The results are close to the experimental values and the errors are, as in the case of G_ij, attributed mainly to the fact that the local elastic constants s_ij in the vicinity of impurity ions are different from the host values. The coefficients F_ij for MgO:V²⁺ crystals, in which the local values are very similar to the bulk values, are also predicted. The results should be, as shown for the coefficients G_ij, closer to the experimental values calculated from the bulk elastic constants. This point remains to be verified.

A semi-empirical embedded-cluster scheme useful for the study of adsorption on metal surfaces, especially charge transfer between adsorbates and substrates, is presented. To avoid the incorrect description of the high-lying antibonding s and p states of metals by the usual cluster calculation, free-electron-like states bounded in the normal direction are introduced in the basis function to describe the s and p states of the transition- or noble-metal substrate, while the d state is represented in the tight-binding approximation. In addition, effects from the dangling d bonds at the boundary are removed by a projection technique. Such a treatment is able to give a reliable local electronic structure of adsorption systems. Application of this scheme to Cs/W(100) and Cs/Cu(100) shows that the calculated results are in close agreement with those by a first-principles linearised augmented plane waves method.

The results of temperature and time (at T=4.2 K) dependent magnetisation (M) measurements on the polycrystalline high-T_c superconductor La_1.8Sr_0.2CuO₄ at different magnetic fields H, 4 kOe<or=H<or=35 kOe, are reported. The diamagnetic shielding and Meissner effect are studied. Like other authors, the authors found a glass-like time dependence of the shielding effect. A new phenomenon is found in the Meissner effect: the inversion of the sign of the magnetisation from negative to positive during a time in which other conditions are unchanged. The differential susceptibility remains diamagnetic in the state with positive M. It is shown that there exists a temperature T₀<T_c such that when T<T₀ the usual mixed state with M<0 for the field cooled regime is metastable and evolves rapidly to a new metastable mixed state with M>0 and dM/dH<0. Such a phenomenon is not observed in traditional type II superconductors.

The crystal structure of Bi₂CuO₄ has been reinvestigated and its magnetic structure solved by neutron powder diffraction. The crystal structure, according to earlier work, is perfectly described in the space group P4/ncc. Contrary to the dimeric magnetic behaviour suggested in a recent paper, the title compound becomes long-range antiferromagnetically ordered below 50 K. The ground state has C_z symmetry and the crystallographic magnetic group is P4/n'c'c'. The magnetic moment of copper atoms at 1.5 K is 0.93(6) mu _B, which is very close to the saturation moment of spin-only Cu(II)(d⁹).

A formalism for the temperature-dependent dynamical magnetic spin susceptibility of d-band metals is developed in the framework of the generalised non-local model potential approach. Using the Shaw-Harrison model wavefunction transformation the magnetic spin susceptibility is separated into two parts: one is caused by the non-nodal character of the model wavefunction (i.e. the free-electron contribution) and the other by the depletion hole associated with the ion core. The latter is responsible for the Bloch character of conduction electrons and includes the s-d hybridisation effects in the d-band metals. The depletion hole contribution is calculated by two different approaches: the first is exact depletion hole approach and the second the averaged depletion hole approach. The calculations are performed for the spin susceptibility of V metal. The values of the depletion hole contribution in both approaches are found to be nearly the same quantitatively as well as qualitatively and are about 46% of the free-electron value. The spin susceptibility decreases and the peaks are broadened with increase in the temperature. The calculated and the experimental values of bulk magnetic spin susceptibility show reasonably good agreement and emphasise the importance of the depletion hole contribution.

The authors propose a new method for the simulation of spin dynamics based on the Langevin equations and apply it to an anisotropic ferromagnet on the square lattice. The authors method reproduces the temperature dependence of the magnetisation in the Monte Carlo method. The dynamical transverse susceptibility G(q, omega ) is calculated and the dispersion relation is obtained from the peak position of Im G(q, omega ). The dispersion relation at the lowest temperature follows the results in the linearised spin-wave theory. They also investigate the temperature dependence of the dispersion relation.

The author investigates the percolation threshold, rho _p, of three-dimensional continuum attractive and repulsive square-well, triangular-well and parabolic-well fluids by Monte Carlo computer simulation. He finds that in the hard-core limit an attractive well decreases rho _p below the high-temperature limiting value. In contrast a hard shoulder potential produces the opposite trend. The author quantifies the role of the range and shape of the well potential in determining the value of the percolation threshold. He examines the shapes of all the clusters at the percolation threshold, resolved as a function of the number of particles in a cluster, s. The asphericity parameter, A₃, describing the instantaneous shape of the cluster, decays slowly from unity, achieving approximately=0.3 by s approximately=70, statistically indistinguishable from the estimated universal value of 0.312 for percolation clusters.

A one-dimensional system, with L sites and periodic boundary conditions, is considered. N divalent atoms occupy sites. The atoms have two orbitals, electron hopping is allowed between nearest neighbour atoms, and intra-atomic electron correlation is taken into account. The ground and first excited many-body states are estimated using a basis, appropriate to weak interatomic coupling, in which multiple nearest neighbour resonating valence bond excitations, appropriately symmetrised, are included. Ordered and disordered systems are considered, for large L.

Raman spectra have been measured for oriented single crystals of NaClO₃ II from 77 K to the melting temperature, for the molten salt and the high temperature phase, NaClO₃ I. The results for NaClO₃ II were in excellent agreement with unit-cell group analysis. Apparent anomalies associated with the presence of weak bands due to naturally abundant ³⁷Cl and ¹⁸O substituted ions have been accounted for on the basis of intermolecular coupling models. The unusual high temperature, ambient pressure phase transition for NaClO₃ has been characterised by spectroscopic and thermal methods. The novel feature about the high temperature phase is the fact that it may only be reached by cooling the melt and it transforms back to the room temperature phase at about 510 K as the sample is cooled. The longitudinal optical mode at about 1030 cm^-1 which was characteristic of the acentric room temperature phase (P2₁3, T⁴) was conspicuously absent in the Raman spectrum of the high temperature phase. The Raman spectrum of the high temperature phase indicated that the crystal has a centric structure but the similarity of many bands with the spectrum of the room temperature phase suggested that the phase transition involved only small changes in atomic positions. The structure remained ordered but there was evidence of considerable band broadening due to increased thermal motion. The Raman results are consistent with the diffraction studies of Meyer and Gasperin (1973) who reported a monoclinic structure similar to room temperature KClO₃, space group P2₁/a. Measurements of the heats of fusion and transition by differential scanning calorimetry gave values of Delta H_f=21.7 kJ mol^-1, Delta S_f=40.6 J mol^-1, Delta H_t=4.1 kJ mol^-1, Delta S_t=8.2 J mol^-1 K^-1. The relatively small values of the heat and entropy of transition compared to those of fusion also indicated that the high temperature phase was not disordered.

The concept of random walk of an excitation within a Gaussian density of states (DOS) is applied to treat diffusion and viscous motion of glass-forming elements controlled by the random potential established upon supercooling a melt. It relates the super-Arrhenius-type temperature dependence observed for viscosity and related properties at T>T_g (the glass transition temperature) to the energetic relaxation of the glass elements within the DOS. The resulting relaxation pattern implies that the system must become non-ergodic at the temperature where the time required to relax to dynamic equilibrium exceeds the experimental time-scale. The model is able to explain quantitatively (i) eta (T) data in the temperature range T_c>T>or=T_g (T_c being a critical temperature above which collective effects, tractable within the mode-coupling concept, become important), (ii) the dependence of T_g on cooling rate and (iii) the Arrhenius-type T dependence of molecular motion below T_g, and qualitatively (iv) the occurrence of physical aging and (v) non-exponential relaxation patterns.

Experimental data on ordinary diffusion coefficients and thermal diffusion factors for binary liquid systems containing hexane, heptane, benzene, toluene and carbon tetrachloride are derived from steady-state and non-steady-state separation measurements in a thermal diffusion column. The data extraction method uses a time evolution theory previously developed by the authors that accounts for the forgotten effect. Comparison of the diffusion data with those of the literature provides a test of the derived thermal diffusion factors.

A model (pseudo)potential method has proven to be an interesting technique for calculating electronic properties in the framework of the Ziman formula for pure liquid metals and the Faber-Ziman expression for liquid alloys. First-principles model potentials are non-local and their parameters are energy dependent. In the alloy one must take into account an energy dependent effective mass, a depletion hole and the Fermi energy core shift which is not known for alloys. Furthermore, the thermopower explicitly includes an energy dependent contribution. The first-principles model (pseudo)potential of Bachelet, Hamann and Schluter (BHS) is energy independent and avoids the necessity of making these corrections (see Phys. Per. B, vol.26, p.4199, 1982). It has been used to calculate analytically the corresponding form factor. The resistivity and the thermopower of pure germanium and gallium and of the germanium-gallium alloy have been determined by using the Faber-Ziman formalism. Both hard sphere and experimental structure factors have been used (when available). The agreement between experimental and theoretical properties can be considered as good and confirms that the BHS model potential is adequate for describing the electronic properties of liquid alloys.

The authors have studied the Raman spectra of a network glass former that derives from GeS₂ by interrupting the sulphur bridges by two halogens, concentrating on GeSBr₂ where on average two of the four bridges per GeS_4/2 tetrahedron are interrupted. The temperature range was 520-230 K (glass transition temperature T_g=250 K). At a resolution of 2 cm^-1, a rich structure develops instead of the very broad lines of GeS₂ glass. They concentrate on the lines that derive from the tetrahedral A₁ (breathing) mode. In order to analyse them, the authors use a program for eigenfrequencies and eigenvectors neglecting symmetries (NORCOR of D Christen). Electro-optical parameters are estimated from the bond polarisability model. They consider quasi-molecular subunits of one, two or three tetrahedra with S or Br substituents. Then the fine structure of the Raman spectra can be understood as due to intermediate-range order, viz, coupling of neighbouring tetrahedra, probability of occurrence of various possible subunits and their geometrical configuration; about 100 subunits are included. A very satisfactory simulation of the Raman spectrum for the said modes is achieved. Edge-connected bitetrahedra and chair-shaped rings are essentially absent. Force constants show only minor changes.

Soliton dynamics in hydrogen-bonded systems are usually modelled by considering the system to consist of two interacting sublattices: one of protons with a non-linear on-site potential and the other of coupled heavy ions. Using the potential recently proposed by Pnevmatikos (1987-88) but with a linear interaction between the sublattices, the authors propose an alternative model which explains the formation and propagation of ionic and Bjerrum defects in ice and other hydrogen-bonded systems.